gpt2-open-instruct-v1-gsm8k

This model is a fine-tuned version of vicgalle/gpt2-open-instruct-v1 on the gsm8k dataset. It achieves the following results on the evaluation set:

• Loss: 2.3966

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0005
• train_batch_size: 16
• eval_batch_size: 1
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss
No log	1.0	468	2.5579
2.859	2.0	936	2.5018
2.6455	3.0	1404	2.4752
2.6025	4.0	1872	2.4590
2.5777	5.0	2340	2.4473
2.5557	6.0	2808	2.4388
2.538	7.0	3276	2.4309
2.5246	8.0	3744	2.4236
2.514	9.0	4212	2.4186
2.5059	10.0	4680	2.4159
2.4944	11.0	5148	2.4107
2.4874	12.0	5616	2.4078
2.4862	13.0	6084	2.4053
2.475	14.0	6552	2.4027
2.4716	15.0	7020	2.4008
2.4716	16.0	7488	2.3995
2.4704	17.0	7956	2.3985
2.4648	18.0	8424	2.3973
2.4634	19.0	8892	2.3968
2.459	20.0	9360	2.3966

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1+cu118
• Datasets 2.13.1
• Tokenizers 0.13.3